Automatic Git Logging: Captures commits via global git hooks
LLM-Powered Summarization: Compresses commit info into concise standup messages
Multiple LLM Providers: Support for Anthropic Claude and Ollama
Manual Logging: Add work items manually
Daily Standup Generation: Generate formatted standup messages from your work logs
Per-Repository Control: Opt-out specific repositories from tracking

Installation

./install.sh

This will:

Create a Python virtual environment at ~/.jrnl/venv
Install dependencies
Initialize the SQLite database
Create default configuration
Optionally install global git hooks
Install the jrnl CLI command to ~/.local/bin

Configuration

After installation, configure your LLM provider:

# For Anthropic Claude
jrnl config set anthropic api_key YOUR_API_KEY

# Or for Ollama (local)
jrnl config set-provider ollama

Configuration file: ~/.jrnl/config.json

Usage

Manual Logging

jrnl new -m "Had meeting with customer about the service"

View Logs

# View recent logs
jrnl logs

# View logs from last N days
jrnl logs --days 3

Generate Daily Standup

# Generate standup from logs since last daily
jrnl daily

# Generate standup covering multiple days
jrnl daily --days 2

# Regenerate today's standup
jrnl daily --regenerate

Configuration Management

# Show current configuration
jrnl config

# Switch LLM provider
jrnl config set-provider ollama

# Set provider-specific settings
jrnl config set anthropic model claude-3-5-sonnet-20241022
jrnl config set anthropic max_tokens_daily 1000

Repository Exclusion

# Exclude current repository from tracking
jrnl config exclude-current

# Exclude specific repository
jrnl config exclude /path/to/repo

# Re-enable repository
jrnl config include /path/to/repo

How It Works

Git Hooks: When you make a commit, the post-commit hook captures the commit message and diff
LLM Compression: The commit info is processed through your chosen LLM to create a concise summary
Database Storage: Logs are stored in SQLite at ~/.jrnl/jrnl.db
Daily Generation: When you run jrnl daily, all logs since your last daily are sent to the LLM to generate a formatted standup message

Project Structure

~/.jrnl/                  # Application directory
├── config.json           # Configuration
├── jrnl.db              # SQLite database
├── venv/                # Python virtual environment
└── logs/                # Application logs

Uninstallation

./uninstall.sh

Or from anywhere:

jrnl uninstall

Requirements

Python 3.8+
Git
Anthropic API key (for Claude) or Ollama installation (for local models)

License

MIT

Development

The project follows a modular structure:

jrnl/commands/ - CLI command implementations
jrnl/database/ - Database layer with SQLite
jrnl/llm_providers/ - LLM provider abstractions
jrnl/git_integration/ - Git hook and commit processing
jrnl/utils/ - Utility functions

To run from source:

python3 -m jrnl --help